Researchers at the Hadassah Medical Organization have achieved a major breakthrough in personalized cancer care utilizing artificial intelligence.
Cancer development is driven by mutations in DNA that leave behind so-called mutational signatures –– unique patterns that act as fingerprints, revealing the underlying causes of the cancer. Some of these fingerprints have clinical implications for treatment plans.
A new neural embedding model developed at the Hadassah Cancer Research Institute, the MESiCA model, uses machine learning, a kind of artificial intelligence, as well as natural language processing, to detect dominant mutational signatures in targeted gene panels using only a few mutations, making it ideal for everyday clinical use. The research is documented in the peer-reviewed publication, Cell Reports Medicine.
“MESiCA represents a major leap forward in personalized medicine,” said Dr. Michal Lotem, head of the Hadassah Cancer Research Institute. “This innovative model offers a practical tool for oncologists, enabling them to make more informed decisions, ultimately leading to improved patient outcomes.”
“MESiCA has been validated in over 60,000 cancer samples, revealing crucial signatures linked to better survival rates and treatment responses,” said Dr. Shai Rosenberg, senior neuro-oncologist and head of Hadassah’s Laboratory for Cancer Computational Biology and the study’s principal investigator.
Read More
Organoids Help Personalize Treatment for Cancer Patients at Hadassah
Breakthrough Research at Hadassah Cancer Institute Pinpoints Harmful Mutations